NEW 117688
REGRESSION (r151021): http/tests/security/video-poster-cross-origin-crash.html is flakily crashing 
https://bugs.webkit.org/show_bug.cgi?id=117688
Summary REGRESSION (r151021): http/tests/security/video-poster-cross-origin-crash.htm...
Zan Dobersek
Reported 2013-06-17 00:07:42 PDT
The http/tests/security/video-poster-cross-origin-crash.html layout test is occasionally crashing on the GTK WK1 builders, most likely after r151021. http://trac.webkit.org/changeset/151021 The crash is occurring in WebCore::MediaPlayer::referrer. Crash log for DumpRenderTree (pid 32194): ... [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Core was generated by `/home/slave/webkitgtk/gtk-linux-64-debug-wk1/build/WebKitBuild/Debug/Programs/D'. Program terminated with signal 11, Segmentation fault. #0 0x00002b56095ab018 in WebCore::MediaPlayer::referrer (this=0x2b5664095090) at ../../Source/WebCore/platform/graphics/MediaPlayer.cpp:1068 1068 return m_mediaPlayerClient->mediaPlayerReferrer(); ... Thread 1 (Thread 0x2b5617b00460 (LWP 32194)): #0 0x00002b56095ab018 in WebCore::MediaPlayer::referrer (this=0x2b5664095090) at ../../Source/WebCore/platform/graphics/MediaPlayer.cpp:1068 #1 0x00002b56095a5cb8 in webKitWebSrcStart (src=0x15b4360) at ../../Source/WebCore/platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:487 #2 0x00002b560d3c4cf2 in g_timeout_dispatch () from /home/slave/webkitgtk/gtk-linux-64-debug-wk1/build/WebKitBuild/Dependencies/Root/lib64/libglib-2.0.so.0 #3 0x00002b560d3c2f95 in g_main_dispatch () from /home/slave/webkitgtk/gtk-linux-64-debug-wk1/build/WebKitBuild/Dependencies/Root/lib64/libglib-2.0.so.0 #4 0x00002b560d3c3cfa in g_main_context_dispatch () from /home/slave/webkitgtk/gtk-linux-64-debug-wk1/build/WebKitBuild/Dependencies/Root/lib64/libglib-2.0.so.0 #5 0x00002b560d3c3eea in g_main_context_iterate () from /home/slave/webkitgtk/gtk-linux-64-debug-wk1/build/WebKitBuild/Dependencies/Root/lib64/libglib-2.0.so.0 #6 0x00002b560d3c431a in g_main_loop_run () from /home/slave/webkitgtk/gtk-linux-64-debug-wk1/build/WebKitBuild/Dependencies/Root/lib64/libglib-2.0.so.0 #7 0x00002b560c583c1d in gtk_main () from /home/slave/webkitgtk/gtk-linux-64-debug-wk1/build/WebKitBuild/Dependencies/Root/lib64/libgtk-3.so.0 #8 0x00000000004a8e75 in runTest (inputLine=...) at ../../Tools/DumpRenderTree/gtk/DumpRenderTree.cpp:769 #9 0x00000000004a8540 in runTestingServerLoop () at ../../Tools/DumpRenderTree/gtk/DumpRenderTree.cpp:552 #10 0x00000000004ab76c in main (argc=2, argv=0x7fffe716aba8) at ../../Tools/DumpRenderTree/gtk/DumpRenderTree.cpp:1511
Attachments
Add attachment proposed patch, testcase, etc.
Philippe Normand
Comment 1 2013-06-17 00:27:51 PDT
Spooky stack-trace!
Zan Dobersek
Comment 2 2013-06-17 04:32:07 PDT
I guess it's a threading problem. I believe the call should be made on the main thread? And the crash occurs because it isn't on the main thread? If so, this should hopefully be easy enough to fix by using WTF::callOnMainThread or something similar.
Andre Moreira Magalhaes
Comment 3 2013-06-17 10:19:59 PDT
(In reply to comment #2) > I guess it's a threading problem. I believe the call should be made on the main thread? And the crash occurs because it isn't on the main thread? > > If so, this should hopefully be easy enough to fix by using WTF::callOnMainThread or something similar. This is strange, the changes I did on bug #115352 should make sure webKitWebSrcStart is called on the main thread with g_timeout_add as you can see from the backtrace.
Note You need to log in before you can comment on or make changes to this bug.